Digitalization is transforming and opening up exciting new opportunities in the power industry – no more so than additive manufacturing, commonly known as 3D printing. Tildy Bayar visits a first-of-its-kind workshop for the development, manufacturing and repair of power generation components in metal through 3D printing
Operating the EOS 3D printer
Additive manufacturing, or laser sintering, or selective laser melting – more commonly referred to as 3D printing – has the potential to effect fundamental changes in the way power plant components are designed, manufactured and distributed.
The leading maker of 3D printing equipment, EOS e-Manufacturing Solutions, notes that ‘additive manufacturing’ is the most accurate term, contrasting it with traditional manufacturing techniques which it calls ‘material removal’ (and which thus might be thought of as ‘subtractive manufacturing’).
Rather than beginning with a solid block of plastic or metal and hewing a shape from it, or welding different subcomponents together, as in traditional techniques, additive manufacturing builds components from scratch through the application of many micro-thin layers of powderized material, applied sequentially and melted together with a precision laser. These materials can be plastics, metals or composites.
While the technique has been in use for 10 years for rapid prototyping purposes in metal, it is now being increasingly used in serial production, where it is already proving transformative. On a recent trip to Siemens’ new additive manufacturing facility in Finspang, Sweden, the technology and, perhaps, the future of power plant component production was on display.
From dreams to reality
Siemens’ Industrial Turbomachinery facility in Finspang is one of three additive manufacturing ‘hubs’ the company maintains, with an existing facility in Berlin and another in the works in the UK, at the headquarters of recently-acquired advanced manufacturing firm Material Solutions. The Finspang facility has been a turbine manufacturing plant since 1913 and had a number of owners, including ABB, before being acquired by Siemens in 2003. It now produces and packages the company’s SGT-500, SGT-600, SGT-700, SGT-750 and SGT-800 model industrial gas turbines, as well as the Industrial Trent (up to 66 MW) aero-derivative gas turbine.
Jenny Larfeldt, Combustion Engineer
Credit: Tildy Bayar
The additive manufacturing department was added to the facility in 2009 and officially opened this year. It features a number of EOS 3D printers which work in conjunction with a proprietary computer-aided design (CAD) program. The first piece printed through its process was a model in metal of the company’s local headquarters, Finspang House.
As Thorbjoern Fors, CEO of the company’s Distributed Generation Services business, puts it, the department’s focus is driven by the idea that “if you can dream it, you can print it”. He says the technology and innovations 3D printing can give rise to could transform both production and service, enabling the manufacturing process to become design-driven and allowing for almost unlimited innovation in design, materials and structures, while also simplifying and speeding up repairs.
In addition, Dr Vladimir Navrotsky, CTO of the Distributed Generation Services business and Siemens’ Innovator of the Year for 2015, notes that “with this technology there is no scale effect – we could make one component or 1000 and the cost would be the same, unlike classical production where bigger volume equals lower cost.”
According to Navrotsky, additive manufacturing offers a number of benefits for the component repair process, including reduced lead time, fewer process steps (for example, no casting is involved), savings on materials, elimination of tools, and on-demand (or “instant decentralized”) production.
In a 2013 example of a repair undertaken through additive manufacturing, the first commercial product produced by the Finspang facility was a spare part, a burner tip for the SGT-800 turbine. Rather than manufacturing a new head, the old head was cut off, and a new head was printed directly onto the burner instead of needing to be welded on.
An SGT-1000F turbine at a gas-fired power plant in Brno in the Czech Republic has been running since June with 3D printed spare parts. During the last scheduled inspection, three of its 24 burners were equipped with Siemens’ first printed burner heads.
“This is the first time this spare-part-on-demand process was ever done,” says Fors, adding that in the traditional process, re-manufacturing the burner head alone would have taken six months; with additive manufacturing, the entire process was accomplished in a few weeks.
Ultimately, Fors says, the aim is to locate repair services in a ‘spare parts on demand’ additive manufacturing centre close to the customer’s location. Navrotsky adds that a “network of printers connected through a secure line through which we send data files” is envisioned for the future, enabling on-demand printing in multiple locations. He says this potential is still “a few years away”, but that Siemens has identified potential sites for such facilities. When pressed, Fors says these centres likely “won’t be in Europe”, as the company already has three such facilities here.
One way additive manufacturing technology could prove transformative is in turning the focus of manufacturing toward the design process, so that – as EOS puts it – “design determines production, and not the other way around”. In other words, current limitations on what can be produced, in terms of both materials and possible structures, could change within a digitally integrated design-production approach, and these changes could ultimately transform power plant components as we know them.
Fors says that, with the ability to print trial designs, many costly aspects of the testing and validation process can be eliminated, while engineers can try out more possible models because “it’s now ok to fail”. In the conventional product development process, testing represents the final validation step and takes place only after years of work. This process, with a defined series of stages, fosters a conservative development approach, Fors says, within which development cycles are long and costly, and developers tend to stick to “moderate” aims.
But with testing integrated within the development process, he says, the cycle can be shortened, the entire process accelerated, and the end goals can become more ambitious and even radical, as testing becomes a low-risk proposition. He says he encourages his team to “go ahead and fail” in order to increase the scope of what can be imagined.
Navrotsky echoes this point, noting that designers today “can do 10 or 15 different versions [of a component design] within three to five years. But now that a design is allowed to fail, we are shrinking that design time to two years or two months, but the validation in the engine still takes three to four years. So the next challenge is how to shrink the validation period. This means our simulation model will be well-calibrated and its precision level will be much higher.”
3D printed burner head
The “go ahead and fail” approach has already resulted in new product innovations. Navrotsky describes a gas turbine blade design that was “previously impossible” to produce given its complexity, but can be printed additively. Currently in the development-and-design phase, the blade could ultimately feature an internal lattice structure designed to reduce its weight, allow for thinner walls, and offer improved cooling and life extension.
Navrotsky also describes a fuel strainer, previously custom-made, which can now be printed “hole by hole”. Instead of the traditional round holes, he says the 2400 0.3-mm holes in a single strainer can now be triangular, oval or any other shape. And, as the holes are no longer drilled, the manufacturing process is simplified.
Navrotsky is excited about the possibilities offered by additive manufacturing, imagining a future where “we can copy nature, such as the design of human bones” to create more efficient and resilient components. But he notes that “you cannot apply this technology everywhere. You can print from plastic materials, but from metal it’s a little bit limited”. However, since additive manufacturing offers the potential to experiment with combinations of materials through mixing the metal powders, various combinations can be used – such as the nickel-based alloys used for a recently redesigned turbine burner.
The power sector’s high-load, high-temperature bearing materials are “not so easy to handle with additive manufacturing,” Navrotsky says, noting that “our business is one of the most difficult application fields” for the technology and that current applications for 3D printing “may not be for any engine components or auxiliary system – that is too expensive – but we aim to penetrate the market starting from the combustor, burner, vane and blades.”
Indeed, a burner swirler for the SGT-750 turbine, which can only be produced through additive manufacturing due to its filigree channels and free-form surfaces, has been in commercial operation for two years.
Jenny Larfeldt, Combustion Engineer at the Finspang facility, notes that the manufacture of components must be approached not just in material terms, but also in terms of their printability, and says the gas turbine burner is “a good piece to print”. While the traditional model contains 13 different parts, a version can be printed as a single piece. In combination with changes in materials, this can reduce the burner’s weight, which means it will print faster.
Attempts to print burners also led Larfeldt’s team to a design innovation which arose from a limitation inherent in the printing process. Because it would not make sense to print a burner in its traditional form, as the additive manufacturing process cannot work against gravity to print an overhanging powdered metal layer (additive manufacturing “has problems with so-called ‘downskin areas’, or angles on overhanging structures that are not supported by solid metal,” Larfeldt says), it was necessary to redesign it aerodynamically. This design innovation also had the knock-on effect of eliminating the traditional need for welding smaller parts together.
Currently, Larfeldt says, “we can print a full set of burners for the SGT-800 turbine, which takes one week. We can also print a set of eight burners at the same time, each configured for different fuels.”
According to Fors, additive manufacture of an entire burner unit is set to be on a commercial footing “within weeks”. And the company claims these new burners will produce significant savings: for 3D-printed SGT-800 burners, Siemens’ testing has revealed potential fuel savings on the order of €3.2 million per year.
Additive manufacturing also presents some technological challenges in terms of normal stresses on gas turbine materials during operation, Fors says. For example, the thermal loading on an operational gas turbine is close to the melting point of the metals used – in the case of iron, around 1500°C. The centrifugal force affecting the blades is 10,000 times the net weight force, and the blade tips reach sonic velocity. However, Fors says that additive manufacturing also makes it possible to address these challenges through the design process for the first time.
Through expanded, and theoretically unlimited, options for internal and external cooling systems for blades and vanes, Navrotsky says the cooling system can be made more effective, and savings in the cooling air can result in a 0.5-1 per cent efficiency improvement for the engine.
In addition, it is possible to increase coating adhesion for blades and vanes through micro-scale engineered surfaces, and newly developed metal powder alloys can increase component lifetime.
It is exciting to be able to design not only components, but the materials that make them, says Navrotsky. But he notes that with the ability to combine new materials comes the need for new standards. “There is quite strong impact from some chemical elements in the material” on the properties of the eventual product, he says, and thus the range of deviation allowed will differ from traditional metals and metal alloys. While Siemens’ team is currently participating in the development of ISO standards, Navrotsky says standards bodies are “still conservative” and “we need to change mindsets”.
While traditional component manufacturing relies on volume, increasing decentralization in the power sector and growing market volatility are giving rise to new business strategies such as customized mass production. In this regard, flexibility is as crucial for component manufacturers as it is for power plant operators. For Siemens, Fors says additive manufacturing can provide that flexibility, with the ultimate goal being production on demand. An important cost reduction factor associated with this model is that there is no need to stock parts, while the company can also react more quickly to customer requests.
3D printed burner fronts
3D printing in process
Navrotsky offers an example of a customer that needed to replace a water pump impeller for a nuclear power plant. While the component is no longer in production, he says that through scanning, modelling and printing, “we could produce it very quickly”.
“In general,” he says, “what we would like to do is come to customers and not say ‘please buy a new component’. Instead we can say ‘forget about components, forget about the supply chain, we’ll give you an extra one per cent efficiency. This is our new business model.”
“Customers don’t need inventory at the same level anymore, because we can deliver very quickly,” says Fors, adding that Siemens’ statement that additive manufacturing reduces lead time by 50 per cent is “confirmed”.
However, some scaling up is still required. “In order to extend the application and to make more parts printed,” Navrotsky says, “we need to work with the size of the machines to print bigger components. We are addressing this with EOS.”
The world of tomorrow
Andreas Graichen, Group Manager for the Additive Manufacturing Centre of Competence in Finspang, says that rather than today’s “islanded” steps through product development stages, Siemens aims to create “a stream” whereby all of these steps are connected to support a “printer cloud”.
He says the first tests involving controlling the 3D printer through the CAD tool are now taking place, and that eventually he believes the firm will be able to “teach remote printers to heal themselves” without human intervention as well as implement robotic cleaning – meaning that decentralized manufacturing centres, or “vertical printer shops”, can be run with minimal human staff and thus located in remote regions. “The thinking behind this is that we can print from 3500 km away,” he says.
In this near-future world, the value of digital models of power plant components may be greater than that of the components themselves, and Graichen says it is the design that will become the valued intellectual property, rather than its physical realization which can be reproduced anywhere, at any time.
For this reason, cybersecurity will become a central issue for any business involved in additive manufacturing, and Graichen notes that Siemens is already actively addressing the issue.
For now, Fors says, the company is still focused on increasing performance and flexibility and reducing lead time. “We are investing heavily in additive manufacturing,” he says, and while “we haven’t yet seen a significant bottom line effect, it’s more that we bring added value to our customers and more value in external markets.
“Down the road it will have cost benefits as well.”